Impedance-based thermal runaway detection and temperature estimation for single and parallel connected large-format automotive lithium-ion batteries
Jan Schöberl, Julian Schumacher, Raphael Urban, Markus Lienkamp
Abstract
Early thermal runaway detection in battery systems of electric vehicles is required to meet legal requirements and to ensure vehicle occupants’ safety. Whereby impedance-based methods offer the potential to detect thermal runaway at an early stage and simultaneously provide a better-resolved temperature estimation during normal operation. However, many studies considering these methods focus only on the cell level at impedances that do not occur in electric vehicles. Consequently, possible challenges and limitations in the transfer to the system level found in electric vehicles are nearly unexplored. This article presents a methodology for early thermal runaway detection and temperature estimation for large-format lithium-ion batteries with low impedance using a parallel connection, as found in the BMW iX3 (G08). The focus is on a methodology that reduces interference factors at cell impedances below 1 mΩ and its use for temperature estimation and thermal runaway detection for single and parallel connected cells. The method is based on the relative change of the real part whereby cell-specific variations from cell-to-cell, the electrical contact resistance, and the system-related measurement setup can be widely compensated. This ensures estimation errors of less than 1 K for both system levels at homogeneous temperature distribution in a temperature range from -10 to 30 °Celsius. More significant errors can be expected at higher temperatures due to a reduced temperature sensitivity. With inhomogeneous temperature distribution, a slight shift of the temperature estimation towards the warmer cell could be observed in the module with a parallel connection. Highly inhomogeneous temperature distribution also increases uncertainty in temperature estimation and impedes thermal runaway detection. However, extensions of the methodology enable the detection of thermal runaway early on both system levels, significantly increasing battery safety in automotive applications.